A longitudinal hierarchical linear model for estimating school effects and their stability

This paper presents ageneral longitudinal model for estimating school effects and their stability. Previous research on the stability of school performance over successive years has produced inconsistent findings. We argue that the findings have been inconsistent for at least two reasons: researchers have estimated ifferent ypes of school effects, and they have not distinguished between instability due to true changes in school performance and instability due to measurement and sampling error. We describe two different ypes of school effects, each relevant o a different policy audience, and we present a longitudinal model that is capable of separating true changes in school effects from sampling and measurement error. The model also provides a means for estimating the effects of school policies and practices while controlling statisti-cally for the effects of factors exogenous to the schooling system. This paper provides an example of the approach based on data describing two cohorts of students from one Education Authority in Scotland. It concludes with a discussion of the limitations of the model and implications for those collecting indicators of school performance for planning and evaluation purposes. Are school effects table over time? Researchers have spent at least 2 decades examining the relationship between schooling inputs, such as student, teacher, and school resources, and student outcomes, uch as academic achievement, self-concept, and career aspirations. The principal aims of this work have been to identify schools that are exceptional in their performance and to determine whether variation between these schools and those less exceptional can be explained by factors that are manipulable by school policies, such as teacher salaries, type of curriculum, or per-student expenditures ( ee Bridge, Judd, &amp